How we track

Registered trials are pulled automatically every night from ClinicalTrials.gov (API v2). The programme register — which company is pursuing what, by which approach — is maintained by hand from peer-reviewed literature, regulatory filings and company disclosures.

A registered trial proves a sponsor declared a study. It does not prove the study is running, or that it will read out. Phase changes detected automatically are queued for review and only published once a human has confirmed them.

We don't rank treatments by expected efficacy, we don't predict launch dates, and we don't take payment to change a record. Programmes that fail stay in the register, marked discontinued — the history of failures is as informative as the history of successes.

Where we operate in the EU, Directive 2001/83/EC Article 88 prohibits direct-to-consumer advertising of prescription-only medicines, so we carry no commercial links to finasteride, dutasteride or any other prescription drug for European readers. Any paid link, when one exists, is disclosed on the page it appears on.

Every record shows a confidence level. It describes the quality of the source, not how sure we feel. High: confirmed by a trial registry, a regulatory document, or peer-reviewed publication. Medium: strongly supported, but by a secondary source — a sponsor's announcement not yet published, for instance. Low: inferred from the product name, the sponsor or a trial title, with nothing establishing it. Anything promoted out of the discovery queue starts at Low however plausible it looks, until a source is added.

The regulator's own document beats the press release reporting it; a peer-reviewed readout beats the sponsor's announcement of it. Where a claim rests only on a company release, the record says so. Every link is checked to resolve before it is published — two FDA URLs that looked canonical were already dead.

This register is maintained by someone who is not a clinician. That is deliberate, and it defines what the site does: it reports what registries and regulators have published, with the source, and does not interpret any of it medically. It will never tell you whether a treatment is right for you — that is a doctor's question. What it will do is make every claim checkable at its source.

The trial import runs nightly. It can create a trial record and flag a phase change, but it can never publish one: a registered trial proves a sponsor filed paperwork, not that a programme advanced. Nothing an automated step produces is visible to a reader until a person has confirmed it.

We take no funding from any sponsor listed, no payment for placement in the register, and no payment to add, change or remove a record. The site currently carries no advertising and earns no commission. If that changes it will be stated here and on the pages affected before the first euro is taken, not after.

A wrong record on a health site does real harm. If you spot one, tell us — corrections are handled ahead of everything else.